AI Startups Grab 51% of VC
AI startups captured 51% of all venture capital funding from January to October 2025, with investments heavily focused on agentic workflows and infrastructure. This week alone, AI chip startups raised $1.1 billion in VC funding, highlighting continued investor appetite for AI hardware and enterprise solutions.
The surge in AI's share of venture capital represents a significant acceleration from previous years. In 2024, AI-related companies captured nearly a third of global venture funding, a figure that has dramatically increased to over half in 2025. This growth is fueled by a global venture market that saw startup funding climb to $425 billion in 2025, a 30% increase from $328 billion in 2024. A handful of massive deals have heavily skewed the funding landscape. In 2025, five companies alone—OpenAI, Scale AI, Anthropic, Project Prometheus, and xAI—accounted for $84 billion, or 20% of all global venture capital. This follows a trend from late 2024, which saw massive rounds like Databricks' $10 billion raise and multi-billion dollar investments in Anthropic and xAI. The focus on "agentic workflows" refers to AI systems designed to be goal-oriented, capable of reasoning, planning, and executing complex multi-step tasks without direct human supervision. Companies in this space, such as UiPath and Aisera, are developing autonomous AI agents to handle entire business processes in areas like IT services, customer support, and finance. The $1.1 billion raised by chip startups this week was distributed among three key challengers to Nvidia's market dominance. MatX, founded by former Google engineers, secured the largest portion with a $500 million round led by Jane Street and Situational Awareness LP to build an LLM-optimized accelerator. SambaNova Systems raised $350 million from investors including Vista Equity and Intel's investment arm for its next-generation dataflow accelerators. Rounding out the week, Dutch startup Axelera AI raised $250 million in a round led by Innovation Industries to scale up production of its chips designed for low-power edge computing workloads.
